Yeah, Clockwork devices have always been a mixed bag. The hardware is well thought out and excellent. (The Devterm and uConsole suffer from one hardware design issue though, that isn’t a problem so much as a “ahead-of-their-time” sort of thing. Since they were designed to support the CM3 and maintain backwards compatibility with it, compromises were made to keep that and the CM4 and CM5 on those devices have some oddities as a result.) Design and unboxing experience are also excellent. And based on my own experience of dealing with a support issue once, and many posts by others on the forum, Alex’s customer service is amazing and given reasonable proof Clockwork always seems to send replacement parts when there are manufacturing defects (my backplate on the uConsole was milled incorrectly), or other hardware issues like a broken screen, etc.
The downsides are, the website looks very professional but always seems to promise things that aren’t true. The big one, that I’m sure has driven plenty of customers away, is the shipping time for most of the devices before the PicoCalc. Granted, COVID delays were a part of that. But I’m not sure there has ever been a consisten time when the devices shipped in “90 days”, and while the website says “90 business days”, I think a lot of people misread that and don’t understand what it means. Realistically, “90 business days” in China seems to be anywhere from 4-6 months. And I’d hazard to guess that probably 75% of the Devterm and uConsole (especially) orders probably weren’t shipped in under 6 months for anyone outside of mainland China. I’ve seen comments and reviewers online who suppose it’s all a scam, but while no one but Clockwork knows what causes the delays, they do eventually ship the devices and make good on them. But they are definitely not a product for anyone on a tight time schedule who can’t wait 6 months or more, or who refuses to pay and not receive anything for that long. It’s unfortunate, unusual, and certainly raises lots of red flags, but they’ve got a consistently good reputation for the devices, even though they have a consistently bad reputation for shipping times and marketing promises. I’ve also noticed that nothing they claim seems to be an outright lie – but it implies thigns are may not be true. Like supported/tested stuff for the PicoCalc… Those things probably will work eventually, but even if someone at Clockwork actually tested them and got them working, it’s unlikely any of that will ever be shared. Or, based on past devices, it might be shared well after the community has gotten it working.
Documentation, board layouts, and 3D models are another thing that eventually gets shared, but usually after lots of asking, begging, whining, and even sometimes threats from the community (since they are often violating licensing agreements by not sharing that info.) I don’t remember exactly how long it took, but I think it was at least a year before all that info got posted for the uConsole, and I know it took a while for the Devterm too. I remember people were already reverse engineering stuff and building add-on modules for the Devterm without it, and even doing multiple hardware revisions for their add-ons because they couldn’t be sure how things actually worked.
A long time ago, when the Gameshell was new, and Clockwork had more employees (or at least people associated with the company) posting on the forums here, they posted some philosophical information which was interesting and also accurate based on how all their devices have been released. They made it very clear that the Gameshell (and presumably future devices) were meant to be hobby projects for tinkerers. It wasn’t meant to be as much as plug-and-play device as a device to dig into, get your hands dirty, customize, and use as a platform to build the device of your dreams. While it would work out of the box, what it offered was more like a model kit, where the user could focus more on software than having to worry about how to customize or fix the hardware (though hardware mods were encouraged too). Weirdly, their marketing doesn’t seem to reflect that philosophy though, as it seems to imply that these are fully functional devices that support all the stuff advertised, rather than being platforms that have the potential to support those things, and more.
So far, the communities for all these devices have done some really incredible things with them. And that’s probably why Clockwork continues to have a following, at least among those who like to tinker, learn, and enjoy customizing a device as much as (or maybe more than!) actually using the device. In a way, these devices are similar to the Flipper Zero – it was deisgned as sold more as a platform, and the people who actually get use out of it are the folks who tinker. Sure, there are kids with Flipper Zeros who “pwn” iPhones around them with spam, and other such nonsense, and there are folks who use the uConsole only to run RetroPie. But I think the real power lies in tinkering, either to build something you want, need, or are interested in, or even just to learn about stuff by doing, even if that just results in lots of projects along the way and a device that is a neverending unfinished project. There are different kinds of fun, and these devices seem to encourage and support fun by doing.
Doesn’t make the ordering/shipping process any less frustrating though. And it requires tempering your own internalizaiton of the hype when a new device is announced. I find thinking about them as platforms for potential projects is a better way to consider them, than as “finished products”. Though to be fair, I think PicoCalc is probably closer to a finished product than anything since the Gameshell. It could be used out of the box to do PicoMite/MMBasic stuff, MicroPython, and Lisp, even if the versions on the SD might be out of date or not support all the features of those things. 
).